Autonomous targeting without human control and untargeted profiling define the primary red lines in this new oversight framework. The author designed these constraints to withstand pressure from Google Cloud and the Pentagon. It prioritizes auditing that respects operational security. Practitioners can use these mechanisms to prevent loophole exploitation in high-stakes government AI procurement.